Clash of Civilizations: Impact of
Culture on Militarized Interstate
Dispute
Gunes Gokmen y
Bocconi University
Preliminary Draft, September 2011
Abstract
Huntington (1993a, 1993b, 1998, 2000) argued that the fundamental source of con‡ict in the post-Cold War world will not
be primarily ideological or primarily economic, but the great divisions among humankind and the dominating source of con‡ict
will be cultural and religious; as such, the primary axis of con‡ict in the future will be along civilizational lines. To that end,
in addition to confronting several of Huntington’s hypotheses we
scrutinize the impact of culture on militarized interstate disputes
and test whether countries that belong to di¤erent civilizations
tend to be more involved in con‡ict than countries that belong to
the same civilization. We show that over the period of 1816-2001
civilizational dissimilarity in a dyad increases the probability of
con‡ict calculated at the means of the variables by up to 62.8
percentage points. More strikingly, even after controlling for geographic, political, military and economic factors, being part of
di¤erent civilizations in the post-Cold War period brings about
71.2 percentage points higher con‡ict probability than belonging
to the same civilization while it reduces the probability of con‡ict
by 25.7 percentage points during the Cold War.
JEL Classi…cation: D74, N40, N70, Z10.
Keywords: civilizations, clash, con‡ict, culture, militarized
dispute.
Department of Economics, Bocconi University, via Rontgen 1, 20136, Milan,
Italy. E-mail Address: [email protected]
y
I am indebted to Alberto Alesina, Marianna Battaglia, Maristella Botticini, Vincenzo Galasso, Lara Lebedinski, Marinella Leone, Andreas Madestam, Monica Masucci, Annaig Morin, Manasa Patnam, Michele Pelizzari and seminar participants at
Bocconi University for their help and support. Needless to say, all errors are mine.
1
In class and ideological con‡icts, the key question was
"Which side are you on?" and people could and did choose
sides and change sides. In con‡icts between civilizations, the
question is "What are you?" That is a given that cannot be
changed. And as we know, from Bosnia to the Caucasus to
the Sudan, the wrong answer to that question can mean a
bullet in the head. (Samuel P. Huntington, 1993)
1
Introduction
In the summer of 1993 Foreign A¤airs published an article entitled "The
Clash of Civilizations?" by Samuel Huntington, which generated a myriad of discussion, controversy, sympathy and antipathy. The article
posed the question whether con‡icts between civilizations would dominate the future of world politics. De…ning a civilization as the broadest
cultural entity and identity under which people form the highest cultural grouping Huntington (1993a, 1993b, 1998, 2000) claims that in the
post-Cold War era the most fundamental source of con‡ict will be civilizational; therefore, the primary axis of con‡ict in today’s post-Cold
War world will be along cultural and religious lines. Consequently, he
concludes that the greatest threat to world peace are the con‡icts between civilizations.
The objective of this study is to scrutinize the impact of culture at
large on militarized interstate disputes and to confront several of Huntington’s hypotheses from di¤erent aspects. To that end, we conduct
analyses to understand whether countries that belong to di¤erent civilizations tend to be more involved in con‡ict than countries that belong
to the same civilization. We run such tests both on cross-sectional level
and over a panel of years between 1816 and 2001. We also distinguish
between countries involved in con‡ict of any level and those escalated
to a war and question the validity of indicator of war as a dependent
variable. Moreover, we break down our investigation into Cold War
and post-Cold War periods and take a closer look into Huntington’s hypothesis by making comparisons between these two time periods. Our
…ndings suggest that civilizational di¤erences do in fact matter. Over
the entire sample of 1816-2001 we show that civilizational dissimilarity
in a dyad increases the probability of con‡ict calculated at the means of
the variables by up to 62.8 percentage points. More strikingly, once the
economic factors are taken into account, being part of di¤erent civilizations in the post-Cold War period brings about 71.2 percentage points
higher con‡ict probability than belonging to the same civilization while
it reduces the probability of con‡ict by 25.7 percentage points during
2
the Cold War.
Impact of ethnicity, language, history, tradition, social norms, culture and religion at large on institutions and the economy have long been
studied.1 In particular, in the context of con‡ict occurrence, there are
studies on both intrastate and interstate con‡ict.2 Fearon and Laitin
(2003), for example, suggest that after controlling for per capita income,
more ethnically or religiously diverse countries have been no more likely
to experience signi…cant civil violence in post-Cold War period. Instead,
they propose poverty, political instability, rough terrain and large populations as the factors that explain which countries have been at risk
for civil war.3 On the other hand, Spolaore and Wacziarg (2010) and
Martin et al. (2008), among others, study interstate con‡icts. Spolaore
and Wacziarg (2010) examine the theoretical and empirical relationship
between the occurrence of interstate con‡icts and the degree of relatedness between countries. Using genetic proximity as a measure of cultural
traits they show that genetically closer populations are more prone to
go to war with each other as these populations share closer ideal points
and a bigger lot of common problematic issues. Martin et al. (2008),
instead, emphasize the role trade links play in interstate dispute and
show that while strong bilateral trade relations reduce the probability of
con‡ict, good multilateral trade relations with third parties increase it.
More speci…c to the Huntington’s hypothesis, there are as well several studies in the literature that try to tackle Huntington’s thesis from
di¤erent angles. Using data from 1820 to 1989 Henderson (1997, 1998)
analyze the impact of cultural factors on the relationship between joint
democracy and war involvement and …nds that both ethnic and linguistic
1
See, for instance, Mauro (1995), La Porta et al. (1999) and Treisman (2000) on
institutions. Furthermore, Easterly and Levine (1997) and Alesina et al. (2003) are
seminal examples illustrating the e¤ects of ethnic, linguistic and religious fragmentation on the economy through their impact on public policies, infrastructure and
productive public goods. In the latter study, it is also important to note that Alesina
et al. (2003) construct a commonly referred to data set on ethnic, linguistic and
religious fractionalization. As a side note, another index of ethnic and cultural fractionalization commonly used by economists and political scientists alike is by Fearon
(2003).
2
For a review of the recent theoretical literature on con‡ict and appropriation
from an economic perspective; speci…cally, by applying conventional optimization
techniques and game-theoretic tools to study the allocation of resources among competing activities— productive and otherwise appropriative, such as grabbing the
product and wealth of others as well as defending one’s own product and wealth see
Gar…nkel and Skaperdas (2006).
3
Another example of intrastate con‡ict is Garcia-Montalvo and Reynal-Querol
(2002). They argue that index of ethnic polarization predicts civil wars better than
the index of fractionalization and suggest that di¤erent measures of heterogeneity
might be needed depending on the question under scrutiny.
3
similarity have direct association with war, whereas religious similarity
within dyads decreases the likelihood of a war; he also …nds that where
a pair of states share a common democratic political culture it exerts a
con‡ict dampening impact that overrides ethnic, linguistic, or religious
factors. Russett et al. (2000) and Henderson and Tucker (2001) assess
the incidents of militarized interstate disputes between countries during
the periods 1950-92 and 1816-1992, respectively. They …nd that such
traditional realist in‡uences as contiguity, alliances, and relative power,
and liberal in‡uences of joint democracy and interdependence, provide
a much better account of interstate con‡ict and intercivilizational dyads
are less, and not more, con‡ict prone. Chiozza (2002) runs an empirical
test of Huntington’s thesis using Kosimo data between 1946-1997 that
include nonviolent con‡ict and shows that state interactions across the
civilizational divide are not more con‡ict prone.4
Though all of them are valuable in their own respect, one should
acknowledge the limitations of the studies described above. First of all,
coverage of the post-Cold War period is extremely limited. And Huntington exactly puts his argument forward and makes predictions about
the post-Cold War world. Secondly, a direct test of Huntington’s hypothesis based on the classi…cation of civilizations by Huntington is not the
central theme of all of the above studies, but rather an index of ethnic,
cultural or religious similarity is used in some of the studies. Thirdly,
survey data are also put under scrutiny. Though survey data might give
useful insights, they rather re‡ect values and beliefs than actual con‡ict
itself. Given the amount of limitations su¤ered by these studies we see
it necessary to confront Huntington’s hypothesis with more recent and
more detailed data taking the basis of civilizations mapped out by him.
The remainder of the paper proceeds as follows. In Section 2 we
take a look at what the clash of civilizations hypothesis is and give a
brief description. In Section 3 we describe the data set used and the
methodology applied for the analyses. Section 4 presents our results.
Finally, Section 5 gives some concluding remarks.
2
What is the Clash of Civilizations Hypothesis?
In 1993 Samuel Huntington published his article entitled "The Clash
of Civilizations?" in Foreign A¤airs, in response to Francis Fukuyama’s
1992 book, The End of History and the Last Man. His article immediately provoked a lot of controversy and heated debate, and subsequently
became one of the most oft-cited articles in the …eld of international re4
Other related studies are Norris and Inglehart (2002), Mungiu-Pippidi and Mindruta (2002), Bolks and Stoll (2003), Gartzke and Gleditsch (2006) and Jakobsen
and Jakobsen (2010).
4
lations. Put brie‡y, Huntington (1993a, 1993b, 1998, 2000) argued that
the fundamental source of con‡ict in the new world (in the post-Cold
War world) will not be primarily ideological or primarily economic, but
the great divisions among humankind and the dominating source of con‡ict will be cultural; as such, the primary axis of con‡ict in the future
will be along cultural and religious lines.
Huntington takes civilizations as the main unit of his analyses. A
civilization is de…ned as "a cultural entity, the highest cultural grouping
of people and the broadest level of cultural identity people have short
of what distinguishes humans from other species. It is de…ned both
by common objective elements, such as language, history, religion, customs, institutions, and by the subjective self-identi…cation of people.”5
Huntington takes the central de…ning characteristic of a civilization as
its religion; hence, the major civilizations in human history have been
closely identi…ed with the world’s great religions. These civilizations
outlined include the Sinic, Japanese, Hindu, Islamic, Orthodox, Western, Latin American, Buddhist and possibly African civilizations plus
"lone" countries that do not belong to one of the major civilizations.
According to Huntington, inter-civilizational di¤erences stand out in
the fashion individuals comprehend the relations between God and man,
the individual and the group, the citizen and the state, parents and children, husband and wife and also in the weight of importance they put in
matters of responsibility and rights, freedom and authority, and equality and hierarchy. He further claims that these di¤erences are largely
irresolvable; they are the product of centuries and far more fundamental
than di¤erences among political ideologies and political regimes as they
concern the very self-identi…cation of man. People’s identifying themselves with a civilization inevitably implies that they think of themselves
separately from other civilizations and di¤erentiate themselves from the
members of other civilizations. To highlight this point, Huntington argues that identity at any level -personal, tribal, racial, civilizational - can
only be de…ned in relation to an "other", a di¤erent person, tribe, race,
or civilization. This brings about a group identity in the simple form of
"us" and "them" which nurtures clashes with those that are di¤erent.
Huntington (1993, 1998), viewing culture more as a “cause,” suggests that civilizations tend to violence with other civilizations that do
not share their culture, worldview, and values. Such violent tendencies,
he argues, long held in check by the Cold War, have been unleashed and
form the dominant pattern of global con‡ict today and in the future.
One theorem that logically devolves from Huntington’s “cultural realist” rendering of “clashing civilizations” is that the degree of cultural
5
Huntington (1993a), p.23-24.
5
dissimilarity between states should predict to the likelihood of con‡ict
between them. In this view, culturally dissimilar dyads, ceteris paribus,
should be more inclined to con‡ict than culturally similar dyads. As
such, Huntington claims that clashes of civilizations are the greatest
threat to world peace, and that in the post-Cold War world the most
important distinctions among peoples are not ideological, political, or
economic. They are cultural. Although, nation states will remain the
most powerful actors in world a¤airs, world politics at the macro level
are likely to involve con‡icts and shifting power balances of states from
di¤erent civilizations, and at the micro level the most violent, prolonged
and dangerous con‡icts are likely to be between states and groups from
di¤erent civilizations. The clash of civilizations will dominate global
politics and the fault lines between civilizations will be the battle lines
of the future.
As for why civilizations will clash Huntington presents several factors. First, di¤erences between civilizations are basic and fundamental.
Civilizations di¤er from one another by history, language, culture, tradition, and, most importantly, religion. These di¤erences are the product
of centuries and cannot be easily overcome. Second, the increased interaction among peoples of di¤erent civilizations makes the world a smaller
place and raises the consciousness and awareness of di¤erences between
civilizations and commonalities within civilizations. Third, as economic
modernization and social change throughout the world separate people
from local identities, a resurgence of religious identity is replacing diminishing local and state-based identities. Fourth, increased civilizationconsciousness sparks a return-to-the-roots phenomenon in non-Western
states. The elites of non-Western societies are going through a deWesternization and indigenization process. Fifth, cultural di¤erences
and characteristics are less mutable and less easily compromised than
political or economic ones. An example highlights this point. A person can be half-French and half-Arab and a citizen of two countries,
but it is more di¢ cult to be half-Catholic and half-Muslim. Finally,
increased economic regionalization heightens civilization consciousness
and, in turn, common civilization facilitates the expansion of economic
relations. The interaction of these factors has resulted in the increased
salience of civilization membership in global politics. Since civilizational
characteristics are basic and essential, civilizational di¤erences are increasingly likely to generate con‡ict.
To support his thesis Huntington depicts several real world incidents.
For instance, wars such as those following the break up of Yugoslavia,
in Chechnya, and between India and Pakistan were cited as evidence
of inter-civilizational con‡ict. To draw attention to the fact that states
6
treat other states di¤erently depending on whether they belong to a similar civilization or a dissimilar civilization, Huntington exempli…es the
failure of the West to provide meaningful support to the Bosnian Muslims in the Yugoslavian War or to denounce Croat atrocities in the same
way Serb atrocities were denounced. To highlight civilizations running
up in support of similar civilizations, he points out Russia’s unwillingness to join other U.N. Security Council members in getting the Serbs
in Croatia to make peace with the Croatian government, and the o¤er
of Iran and other Muslim nations to provide 18,000 troops to protect
Bosnian Muslims during the Yugoslavian War. Another example comes
from the period of the intensi…cation of the war between Armenians
and Azeris: Turkish and Iranian demands that the Armenians surrender their conquests, the deployment of Turkish troops to the border and
Iranian troops across the Azerbaijan border, and Russia’s warning that
the Iranian action contributes to "escalation of the con‡ict" and "pushes
it to dangerous limits of internationalization." Lastly, Huntington makes
an example of the U.S. bombings of Baghdad, its virtually unanimous
support by Western governments, and its condemnation by almost all
Muslim governments as another example of the West’s "double standard."
3
Data and Methodology
3.1
Measurement of Con‡ict
Our data on con‡ict run between 1816 and 2001.6 The indicator of
con‡ict takes on a value from 0 for no militarized dispute to 5 for high
intensity con‡ict that is de…ned as an inter-state war with more than
1000 total battle deaths. The levels of intensity are classi…ed as follows:
1 = No militarized action, 2 = Threat to use force, 3 = Display of force, 4
= Use of force, and 5 = War. In accordance with the literature, we de…ne
an indicator variable taking a value of 1 if the intensity of militarized
con‡ict is equal to or greater than 3, zero otherwise.7 We treat both
indicator of con‡ict and indicator of war (corresponding to a con‡ict
intensity of 5) as dependent variables and run analyses accordingly.
Between 1816-2001 there has been an upward trend in the number of
militarized con‡ict per year with spikes of World War I and World War
II. We also observe a relative increase starting with early 90’s compared
6
Our con‡ict data come from the Correlates of War Project. Data for con‡ict
are Correlates of War Project, Militarized Interstate Disputes, Version 3.10; which is
described in Ghosn et al. (2004) and Ghosn and Bennett (2003). A data set on war
is also available as Correlates of War Project, 2011 COW Wars, 1816-2007, Version
4.0 (for details see also Sarkees and Wayman (2010)).
7
For an example, see Spolaore and Wacziarg (2010).
7
to rather high but stable levels of con‡ict during Cold War.8 On the
other hand, when we look at what part of the militarized disputes is
within the same civilization and what part is between di¤erent civilizations, we see that 36 percent of all con‡ict between 1816 and 2001 took
place between countries that are part of the same civilization; whereas
64 percent of the con‡ictual relationships were among di¤erent civilizational memberships.9 When we break this analysis down to Cold War
and post-Cold War periods, we observe that there is a bigger percentage
of inter-civilizational militarized con‡ict during Cold War than during
post-Cold War, 65 percent compared to 60 percent; which leads us to a
hint contrary to Huntington’s thesis.10
3.2
Measurement of Civilizations
179 countries are classi…ed as members of various civilizations. As described in Section 2 and in Huntington (1998), these civilizations are
Western, Sinic, Islamic, Hindu, Orthodox, Latin American, African,
Buddhist and "Lone" States. Construction of civilization membership
is based on Huntington (1998). Accordingly, each country is assigned to
a civilization.11
Furthermore, country dyads are formed by pairing each country with
one another; which resulted in 15931 dyads. To indicate civilizational
heterogeneity within a dyad we construct a variable labeled as "Di¤erent Civilizations" , DCij ; denoting whether a pair of countries belong
to di¤erent civilizations. This variable is coded as 1 if in a dyad the
two countries belong to di¤erent civilizations and as 0 if both countries
belong to the same civilization. Out of 15931 country-pairs there are
2875 pairs for which both countries belong to the same civilization and
13056 pairs for which countries belong to di¤erent civilizations.
3.3
Other Variables
Geographic Factors Geographical proximity is considered to be one
of the strongest determinants of war. As one of the measures of geographical proximity territorial contiguity is shown to be a strong predictor of con‡ict (Bremer, 1992). The proximity of interactions is likely
to o¤er both the opportunity and the willingness to engage in con‡ict.
Therefore, we take contiguity as one of our independent variables. Our
contiguity variable takes value one if there is any sort of land or water
8
Interested reader can consult Figure 1A in the Appendix.
See Figure 2A in the Appendix.
10
See …gures 3A and 4A in the Appendix.
11
See Table 1A in the Appendix for the details of country speci…c civilizational
memberships.
9
8
contiguity between two countries in a pair, zero otherwise.12
As in Spolaore and Wacziarg (2010), we also control for additional
geographic distance metrics such as the measure of the great circle (geodesic) distance between the major cities of the countries13 , latitudinal
distance, longitudinal distance and indicators of geographic isolation and
geographic barriers such as number of landlocked countries in a dyad and
the land area of the countries.14
Political Factors Factors which might have contributed to the current state of the institutions and the state of the matters between two
countries go back in history. To control for such historical, political
and institutional links we include a dummy variable for whether a dyad
ever had a colonial relationship, i.e. whether one was a colony of the
other at some point in time. In addition, we have a dummy variable
for whether a pair of countries have had a common colonizer after 1945,
i.e. whether the two countries have been colonized by the same third
country. Furthermore, governing bodies leave their legacy on cultural,
historical, political and institutional ties, and this requires inclusion of
a dummy variable to control for whether two countries have been part
of the same polity.15
Democratic peace argument suggests that democratic countries are
less prone to violence and democracy promotes peace (Levy and Razin,
2004). We measure the extent of democracy using the 21-point institutionalized democracy scale in a modi…ed version of the Polity IV data
where -10 means a hereditary monarchy and +10 a consolidated democracy.16 As in Martin et al. (2008), we use the sum of the democracy
indexes of the two countries in a pair.
Di¤erent legal origins have been shown to have strong implications
for institutional outcomes (La Porta et al., 1999). These institutional
outcomes may, in turn, shape the con‡ict pattern between two countries.
Therefore, we create a dummy variable for whether two countries in a
12
For contiguity data we use Correlates of War Project, Direct Contiguity Data,
1816-2006, Version 3.1 (Stinnett et al., 2002). See also Gochman (1991) for additional
details.
13
See Head and Mayer (2002) for details.
14
These data are compiled by the researchers at the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII). The data are available at
http://www.cepii.fr/anglaisgraph/bdd/distances.htm.
15
These data come as well from CEPII. The data are available at
http://www.cepii.fr/anglaisgraph/bdd/distances.htm.
16
The suggested way of categorization and interpretation of these scores by the
project authors is as follows. Scores from -10 to -6 correspond to "autocracies",
from -5 to +5 to "anocracies" and from +6 to +10 to "democracies". The data are
available at http://www.systemicpeace.org/polity/polity4.htm.
9
pair have di¤erent legal origins. This variable takes value one if the two
countries in a dyad have di¤erent legal origins, zero otherwise 17
Military Factors The idea that an equal-balance of military capabilities deters con‡ict in a contest of arms forces us to control for relative
military capabilities of countries (Russet et al., 2000). Therefore, to
assess the e¤ect of states’military capabilities on the likelihood of con‡ict, we use National Material Capabilities data set. The widely-used
Composite Index of National Capability (CINC) index is based on six
variables in the data set: total population, urban population, iron and
steel production, energy consumption, military personnel, and military
expenditure of all state members.18
Major military actors are expected to be positively associated with
dispute involvement (Chiozza, 2002). Hence, we control for the number
of countries with major power status in a dyad as designated by the
authors of the Correlates of War Project.19
Widespread expectations about allies are that they …ght each other
less as they are already in agreement regarding their concerns of security
and also allied states often have other political and economic interests
in common (Russet et al., 2000). To control for the in‡uence of alliances
on con‡ict, we include a dummy variable for whether a pair of countries
are in some form alliance.20
Furthermore, to control for the continuity of con‡ictual relationships
and the contagion from other disputes we construct two variables. One
accounts for the number of peaceful years between two countries since
the last con‡ict has occurred and the other one takes into account the
number of other wars in the world in year t:
Economic Factors To control for whether income di¤erences have an
e¤ect on the likelihood of con‡ict, as in Spolaore and Wacziarg (2010),
we take into account the absolute value of the log income di¤erences
17
Legal origin indicators (common law, French civil law, German civil law, Scandinavian law, and Socialist law) are from La Porta et al. (1999).
18
We use National Material Capabilites data set Version 4.0 from Correlates of
War Project. For details see also Singer (1987) and Singer et al. (1972).
19
Our data come from State System Membership List Version 2008.1 of Correlates of War Project. The designation of major powers also follows COW criteria and includes Austria-Hungary (1816–1918), Prussia/Germany (1816–1918, 1925–
1945), Russia/USSR (1816–1917, 1922–), France (1816–1940, 1945–), United Kingdom (1816–), Italy (1860–1943), Japan (1895–1945), United States (1898–), and
China (1949–). China, France, the USA, the UK, and the USSR are classi…ed as
major powers since 1945, as are the German Federal Republic and Japan after 1991.
20
Alliances data are Version 3.03 from Correlates of War Project (Gibler, 2009;
Gibler and Sarkees, 2004). These data originally date back to Singer and Small
(1966) and Small and Singer (1969).
10
between two countries in a dyad from 1950 on.21
One would also need to control for trade relations as economic dependence makes countries less daring when it comes to con‡ict involvement.
For instance, Russet et al. 2000 divide the sum of a country’s exports
and imports with its dyadic partner by its GDP to see how much this
bilateral trade relation is economically important. They claim that, as
with the in‡uence of democratic institutions, one expects the likelihood
of a dispute to be primarily in‡uenced by the freedom of action available to the state less constrained from using force. This is the state
with the lower bilateral trade-to-GDP ratio, because it is less dependent economically on trade with the other member of the dyad. On the
other hand, Martin et al. (2008) and Spolaore and Wacziarg (2010) take
into account bilateral openness between a pair of countries and their
multilateral openness with third parties. Following this stream we use
bilateral and multilateral openness together with their interaction with
distance. Bilateral openness is constructed by dividing the trade volume
between a pair of countries by the GDP of each country and then taking
the average. Multilateral openness, on the other hand, is constructed by
dividing the trade volume with third parties of each country in a pair
by its GDP and then taking the average.22
3.4
Descriptive Statistics
We observe in Table 1 the number of con‡icts across and within civilizations and their share in total number of con‡icts between 1816 and 2001.
On-diagonal entries correspond to the con‡icts which have happened
between countries that are members of the same civilization; whereas
o¤-diagonal elements are the con‡icts between countries that belong to
di¤erent civilizations. This table gives us a better understanding of
what part of the con‡icts are inter-civilizational and what part intracivilizational; moreover, we get a sense of what civilizations have more
con‡ictual relationships than the others.
We observe many on-diagonal elements; which tell us that there is
a myriad of intra-civilizational disputes, though the numbers are much
smaller than o¤-diagonal entries. By looking at intra-civilizational disputes the most combat prone civilizations seem to be Western and Islamic civilizations followed by a high degree of con‡ict among Latin
American countries with 9.5, 7.9 and 6.8 percents of all con‡ict, respectively .
21
For income data we use Penn World Tables Version 6.2 available at
http://pwt.econ.upenn.edu/php_site/pwt_index.php.
22
Trade data come from Correlates of War Project, Trade Data Set, Version 2.0.
See also Barbieri et al. (2008, 2009).
11
TABLE 1. Number of Conflicts within and across Civilizations between 1816-2001.
Civilizations
Western
Western
419 (9.5%)
Sinic
Islamic
Sinic
265 (6%)
Islamic
581 (13.1%)
15 (.34%)
351 (7.9%)
Hindu
5 (.11%)
33 (.74%)
65 (1.4%)
Orthodox
530 (12%)
Latin American 326 (7.4%)
Hindu
Orthodox Latin American
African
Buddhist Lone States
127 (2.8%)
90 (2.04%) 247 (5.6%)
3 (.06%)
0
105 (2.3%)
7 (.15%)
8 (.18%)
11 (.24%)
13 (.29%)
302 (6.8%)
African
39 (.88%)
0
53 (1.2%)
5 (.11%)
6 (.13%)
21 (.47%)
Buddhist
25 (.56%)
81 (1.8%)
3 (.06%)
6 (.13%)
5 (.11%)
0
0
78 (1.7%)
Lone States
106 (2.4%)
86 (1.9%)
44 (.99%)
0
82 (1.8%)
41 (.93%)
7 (.15%)
16 (.36%)
193 (4.3%)
4 (.09%)
Source: Author’s own construction
One striking observation from Table 1 is that when we look into
highly con‡ictual inter-civilizational linkages we notice that one side usually involves a Western country. For instance, the …rst four highest number of inter-civilizational con‡icts are Western versus Islamic with 13.1
percent, Western versus Orthodox with 12 percent, Western versus Latin
American with 7.4 percent and Western versus Sinic civilizations with 6
percent, respectively. This pattern of high Western con‡ict proneness is
followed by the con‡icts between Islamic and Orthodox civilizations.
For Cold War and post-Cold War break down of Table 1, see tables
2A and 3A in the Appendix. Important to notice in tables 2A and 3A
is that while the highest number of con‡ict during Cold War period is
between Western and Islamic civilizations, it is between Western and
Orthodox civilizations in the post-Cold War period.
For additional insights at a …rst glance see also Table 4A in the
Appendix which provides summary statistics of all variables.
3.5
Empirical Speci…cation
As a starting point, we follow the existing literature23 by running regressions of a binary indicator of con‡ict on several determinants and
collapse the panel into a cross-section (as in Spolaore and Wacziarg,
2010), in which case our dependent variable becomes a binary indicator
of whether there has ever been a con‡ict between a pair of countries over
the period of 1816 to 2001. Given our main explanatory variable, civilizational dissimilarity, is time invariant this speci…cation seems appropriate.
Therefore the baseline cross-sectional regression is the following:
Cij =
23
0
+
1 DCij
+
2 Xij
+ "ij
For example, Bremer (1992) and Martin et al. (2008).
12
(1)
where Cij is an indicator of con‡ict between a pair of countries and
takes the value 1 if the pair of countries were ever involved in a militarized
dispute, DCij is an indicator of civilizational heterogeneity that takes
value 1 when a pair of countries belong to di¤erent civilizations and 0
otherwise and Xij is a vector of control variables such as geographic
factors and colonial and legal indicators.
Moreover, we apply a second methodology to exploit the full panel
dimension. This way we can make use of time varying dimensions of our
explanatory variables. Accordingly, the baseline panel regression would
be as follows:
Cijt =
0
+
1 DCij
+
2 Xijt
+
ijt
(2)
where Xijt contains all of the aforementioned time-invariant variables
plus time varying variables such as democracy, di¤erences in military
capabilities, how many other wars there are in a certain year, how many
years countries have been at peace with each other for, whether they are
part of an alliance, income di¤erences and trade relations.
Equations (1) and (2) are estimated using probit. Throughout the
paper we report marginal e¤ects of the probit regressions evaluated at
the means of the independent variables and for the sake of readibility we
multiply all of the marginal e¤ects by one hundred in all tables. In addition, we report standardized magnitude of the e¤ect of civilizational dissimilarity, which largely eases interpretation. Standardized magnitude
is the e¤ect of a discrete change from zero to one in di¤erent civilizations
dummy as a percentage of the probability of con‡ict calculated at the
means of the variables.
4
Results
4.1
Cross Sectional Analyses
In this section we present our results concerning militarized clash between states and the role culture plays in con‡ict involvement. We …rst
run cross-sectional analyses. We collapse our panel data of con‡ict into
one cross-sectional variable that takes on value one if a country pair has
ever been involved in militarized dispute from 1816 to 2001, zero otherwise. We use probit models and marginal e¤ects in a probit model
evaluated at the means are presented. There are 178 countries in the
cross-section analyses from which country dyads are constructed.
4.1.1
Entire Sample
In Table 2 we present the regression results covering the entire sample;
whereas in Table 3 we break down our analysis and make Cold War and
13
post-Cold War comparisons.
In Table 2 column (1) we start o¤ with a univariate speci…cation. Although the e¤ect of being in di¤erent civilizations on con‡ict is negative
at …rst, it immediately changes sign in the following columns and becomes positive when we account for other determinants of con‡ict such
as contiguity which is one of the most established determinants of con‡ict (Gleditsch and Singer, 1975; Henderson, 1997). We add geographic
control variables to the regression equation in column (2) and notice that
the coe¢ cient on di¤erent civilizational membership is positive. Moreover, contiguity has a positive e¤ect on con‡ict probability. Contiguity
is always highly signi…cant and the biggest in magnitude among all other
determinants of con‡ict in all of the following regression speci…cations
and reappears as one of the biggest determinants of inter-state con‡ict.
We also control for other geographic measures to account for the distance
between countries and the physical barriers within and across countries.
These include the distance between countries, di¤erences in longitudes
and latitudes, whether there are landlocked countries in a dyad and the
physical size of the countries. As expected, the physical distance and
barriers between countries act as a signi…cant deterrent to clashes, and
hence, it reduces the probability of con‡ict.
In column (3) of Table 2 we take into account political factors that
might a¤ect con‡ict likelihood such as colonial links, whether the countries have been part of the same polity and their legal origins. Our
results suggest that colonial and governmental history play a signi…cant
role and they instigate con‡ict involvement. Moreover, when a pair of
countries have di¤erent legal origins it is less likely that they can …nd a
peaceful solution to their problems.
In column (4) of Table 2, instead, we run a similar regression to that
of column (3) using an indicator of war (con‡icts of level 5 with more
than 1000 battle deaths) as dependent variable rather than con‡ict of
level 3 and above. We observe the positive e¤ect of di¤erent civilizational membership on war involvement and all conclusions regarding
other explanatory variables carry on.
To put the importance of belonging to di¤erent civilizations in perspective we take a look at the standardized magnitudes. In columns
(2) and (3) we observe that being part of di¤erent civilizations increases
the probability of con‡ict by 4.2 to 8.2 percentage points; whereas it
increases the probability of war by 34 percentage points as shown in
column (4).
14
TABLE 2. Cross-Sectional Regressions, probit.
(Dependent variable: dichotomous indicator for whether a country pair
was ever involved in a conflict, 1816-2001).
(1)
univariate
(2)
(3)
(4)
add
add
war
geographic polical
(dependent
specification
factors
factors
variable)
-5.8***
.163
.081
.075
(0.000)
(0.5)
(0.73)
(0.18)
10.84***
9.79***
.781***
(0.00)
(0.00)
(0.00)
-1.95***
-1.80***
-.203***
(0.00)
(0.00)
(0.00)
Log Absolute Dif f erence
-.41***
-.435***
-.104***
in Latitudes
(0.00)
(0.00)
(0.00)
Log Absolute Dif f erence
-.064
-.078
-.021
in Longitudes
(0.56)
(0.4)
(0.4)
Number of Landlocked
-2.3***
-2.30***
-.265***
Countries in the Pair
(0.00)
(0.00)
(0.00)
Log Product of Land
.74***
.728***
.124***
Areas in sq km
(0.00)
(0.00)
(0.00)
7.56***
.403*
(0.00)
(0.06)
Countries w ere or are
4.02***
.54**
the Same Country
(0.00)
(0.02)
Dif f erent Legal Origins
.725***
.00193***
(0.009)
(0.00)
Dif f erent Civilizations
Contiguity
Log Geodesic Distance
Ever in Colonial Relationship
# Obs.
Standardized Magnitude(%)
15753
15309
15309
15309
-132.306
8.297
4.231
34.011
p-values in parentheses; *** signif icant at 1%; ** signif icant at 5%; * signif icant at 10%.
Probit marginal ef f ects are reported in all columns. For dummy variables, marginal ef f ects are f or
discrete changes.from 0 to 1. The standardized magnitude is the ef f ect of a discrete change
f rom 0 to 1 in different civilizations dummy as a percentage of the probability of conf lict at the
means of the variables. All marginal ef f ects are multiplied by 100 f or the sake of readability.
15
4.1.2
Cold War and post-Cold War Comparisons
Although we had, by and large, a …rst pass at investigating the impact
of culture on con‡ict involvement in Table 2, it is far from satisfactory
and does not directly test Huntington’s hypothesis. Next, we present a
break-down analysis of Cold War and post-Cold War periods, which are
to be followed by more detailed panel analysis.
Columns (1) and (2) of Table 3 provide us with a comparison of Cold
War and post-Cold War periods, respectively, using geographic controls
while columns (3) and (4) serve the same purpose using additional political control variables. In columns (1) and (3) the dependent variable
is an indicator of whether a pair of countries have ever been involved in
con‡ict between 1946 and 1991 which is the period that corresponds to
the Cold War; whereas in columns (2) and (4) the dependent variable
is an indicator of whether a pair of countries have ever been involved in
con‡ict between 1992 and 2001.24
In both speci…cations with geographic and political factors, countries
that belong to di¤erent civilizations are more likely to be involved in
militarized dispute in the post-Cold War era than in the Cold War era.
During the Cold War the coe¢ cient is both insigni…cant and small in
magnitude. This result delivers support for Huntington’s hypothesis,
that is civilizational di¤erences are more emphasized since the end of the
Cold War and countries that are part of the same civilization con‡ict
with one another less than the ones that belong to di¤erent civilizations.
Being part of di¤erent civilizations boosts the likelihood of con‡ict by
about 70 percentage points during the post-Cold War compared to a 6.9
to 12.3 percentage-point increase during the Cold War.
In addition, importance of distance and physical barriers in con‡ict
involvement is lesser in the post-Cold War period than in the Cold War
period. This might be due to the di¤erences in the advancement of military technology between two time periods, i.e. because countries have
better, more sophisticated military capabilities in more recent times.
Advanced military technology might render more distant wars easy so
that proximity of countries do not play as big of a role as it used to when
it comes to raiding a target.
Furthermore, notice in columns (3) and (4) that dissimilar legal origins lose their consequence in the post-Cold War era. This might be
the case because when communism collapsed more countries adopted
24
See Table 5A in the Appendix for a reproduction of Table 3 using the indicator
of war as dependent variable. In Table 5A, though positive, coe¢ cients on di¤erent
civilizations are not signi…cant in the post-Cold War period. This might be due to
the very small number of wars in the post-Cold War period. There are only 20 wars
in the post-Cold War period.
16
TABLE 3. Cross-Sectional Regressions, probit.
(Dependent variable: dichotomous indicator for whether a country pair
was ever involved in a conflict, Cold War and post-Cold War Comparison).
(1)
(2)
(3)
(4)
Cold War post-Cold War Cold War post-Cold War
period
Different Civilizations
period
period
period
.104
.173***
.057
.166***
(0.46)
(0.001)
(0.69)
(0.001)
6.001***
1.167***
5.33***
1.08***
(0.00)
(0.00)
(0.00)
(0.00)
-.585***
-.306***
-.51***
-.289***
(0.00)
(0.00)
(0.00)
(0.00)
Log Absolute Difference
-.178***
-.17***
-.185***
-.165***
in Latitudes
(0.002)
(0.00)
(0.001)
(0.00)
Log Absolute Difference
-.237***
-.098***
-.239***
-.096***
(0.00)
(0.00)
(0.00)
(0.00)
-1.2203***
-.131***
-1.21***
-.111**
Countries in the Pair
(0.00)
(0.005)
(0.00)
(0.013)
Log Product of Land
.335***
.126***
.324***
.119***
Areas in sq km
(0.00)
(0.00)
(0.00)
(0.00)
1.99***
.302
(0.00)
(0.12)
-.225
-.203***
Contiguity
Log Geodesic Distance
in Longitudes
Number of Landlocked
Ever in Colonial Relationship
Common Colonizer
(0.27)
(0.002)
Countries were or are
2.69***
.209
the Same Country
(0.00)
(0.2)
Different Legal Origins
.291**
-.0086
(0.019)
(0.8)
# Obs.
15309
15309
15309
15309
Standardized Magnitude(%)
12.323
69.626
6.974
70.152
p-values in parentheses; *** significant at 1%; ** significant at 5%; * significant at 10%. Probit
marginal effects are reported in all columns. For dummy variables, marginal effects are for
discrete changes from 0 to 1. The standardized magnitude is the effect of a discrete change
from 0 to 1 in different civilizations dummy as a percentage of the probability of conflict at the
means of the variables. All marginal effects are multiplied by 100 for readability.
17
democracy and this fact washes out the legal di¤erences between countries when both are democratic.25
4.2
4.2.1
Panel Analyses
Entire Sample
Now, we turn to our panel regressions. We have an indicator of con‡ict
for the years 1816 to 2001; hence this gives us a good coverage of the
post-Cold War period as well as a high number of observations. Table
4 reports the results covering 1816 to 2001. Our …ndings from crosssectional regressions are also con…rmed in a panel setting.
In columns (2) and (3) of Table 4 we have the same set of regressors
as in columns (2) and (3) of Table 2, namely geographic and political
determinants of con‡ict, but this time we make use of panel dimension.
In columns (2) and (3) we show that if a pair of countries belong to
di¤erent civilizations they have 55.2 to 62.8 percentage points higher
chances of con‡ict than if they were to be part of the same civilization.
In column (4) we bring the democratic peace argument into the picture and control for the sum of democracy indexes of the two countries
as in Martin et al. (2008).26 As expected, democracy promotes peace.
Our variable on civilizational dissimilarity is still positively signi…cant
and di¤erential civilizational membership in a dyad brings about 41.5
percentage points higher likelihood of con‡ict.
Now we turn to Table 5. Table 5 includes all of the previously discussed geographic and political factors including democracy, but they
are not reported due to space constraints. Instead, we report the additional variables of military and economic factors. Di¤erent civilizational
membership still appears positive and signi…cant and increases the probability of con‡ict by 7.7 to 24.4 percentage points.
In column (1) of Table 5 we add military factors such as number of
major military powers in a dyad, log of absolute di¤erences in military
capabilities of the two countries, whether the pair of countries are part
of an alliance, number of other wars fought in the same year and the
number of years the two countries spent at peace with each other to
the previously present geographic and political factors. Civilizational
dissimilarity is positive and signi…cant. Moreover, notice that big players
in the world scene are more con‡ict prone. If a pair of countries are
part of an alliance they …ght less, but they are negatively a¤ected by
the other wars in the world. It is important to highlight that peace
25
The correlation between di¤erent legal origins dummy and the sum of democracy
indexes variable is about .10.
26
For a discussion on democratic peace argument see Henderson (1997) and Levy
and Razin (2004).
18
TABLE 4. Panel Regressions, probit.
(Dependent variable: dichotomous indicator of conflict between 1816-2001)
(1)
(2)
univariate
(3)
geographic political
(4)
control
specification
factors
factors
democracy
-.8043***
.131***
.1106***
.096***
(0.00)
(0.00)
(0.00)
(0.00)
1.115***
1.065***
1.31***
(0.00)
(0.00)
(0.00)
-.149***
-.144***
-.117***
(0.00)
(0.00)
(0.00)
-.078***
-.074***
-.089***
(0.00)
(0.00)
(0.00)
-.0485***
-.048***
-.065***
(0.00)
(0.00)
(0.00)
-.214***
-.19***
-.223***
Countri es i n the Pair
(0.00)
(0.00)
(0.00)
Log Product of Land
.0874***
.084***
.09***
(0.00)
(0.00)
(0.00)
.221***
.344***
(0.00)
(0.00)
-.063***
-.093***
(0.00)
(0.00)
.131***
.086***
(0.00)
(0.004)
.121***
.1406***
(0.00)
(0.00)
Di fferent Ci vi l i zati ons
Conti gui ty
Log Geodesi c Di stance
Log Absol ute Di fference
i n Lati tudes
Log Absol ute Di fference
i n Longi tudes
Num ber of Landl ocked
Areas i n sq km
Ever i n Col oni al Rel ati onshi p
Com m on Col oni zer
Countri es were or are
the Sam e Country
Di fferent Legal Ori gi ns
-.0026***
Sum of Dem ocracy Indexes
(0.00)
# of Obs.
Standardized Magnitude(%)
590337
583546
583546
488085
-114.749
62.889
55.201
41.555
Robust p-values in parentheses; *** signif icant at 1%; ** signif icant at 5%; * signif icant at 10%.
Probit marginal ef f ects are reported in all columns. For dummy variables, marginal ef f ects are
f or discrete changes f rom 0 to 1. The standardized magnitude is the ef f ect of a discrete
change f rom 0 to 1 in dif f erent civilizations dummy as a percentage of the probability of conf lict
at the means of the variables. A ll marginal ef f ects are multiplied by 100 f or readability.
19
TABLE 5. Panel Regressions, probit.(Dependent variable: dichotomous indicator of conflict between 1816-2001)
(1)
(2)
(3)
military factors income difference trade relations
Different Civilizations
(4)
add post-Cold War Dummy
.0171***
.00931**
.00307
-.00593
(0.00)
(0.04)
(0.55)
(0.33)
.1134***
.0607***
.0457***
.04348***
(0.00)
(0.00)
(0.00)
(0.00)
.0155***
.0082***
.0052***
.00538***
(0.00)
(0.00)
(0.009)
(0.006)
-.0179***
.000723
-.00333
-.00315
(0.00)
(0.9)
(0.59)
(0.6)
.00158***
.000199**
.000277**
.000252**
(0.00)
(0.05)
(0.014)
(0.032)
-.00588***
-.00344***
-.00352***
-.00354***
(0.00)
(0.00)
(0.00)
(0.00)
-.0065***
-.00397**
-.00427***
(0.00)
(0.013)
(0.006)
-.0449***
-.04385***
Openness, t-4
(0.00)
(0.00)
Log Multilateral
.04142
.04777*
Openness, t-4
(0.15)
(0.10)
Log Distance×Log
.00574***
.00567***
Bilateral Openness
(0.00)
(0.00)
Log Distance×Log
-.00665*
-.00755**
(0.07)
(0.04)
Number of Major
Pow ers in the Pair
Log Absolute Difference
in Military Capabilities
Alliance Dummy
Number of Other
Wars in Year t
Number of
Peaceful Years
Absolute Difference
in Log per capita Income
Log Bilateral
Multilateral Openness
Post-Cold War Dummy
-.01657**
(0.03)
Different Civilizations ×
.03994***
Post-Cold War Dummy
(0.001)
# of Obs.
Standardized Magnitude(%)
487276
217188
149646
149646
24.47
22.785
7.736
-14.99 — 86.014
Robust p-values in parentheses; *** significant at 1%; ** significant at 5%; * significant at 10%. Probit marginal effects are reported in all columns.
For dummy variables, marginal effects are for discrete changes from 0 to 1. T he standardized magnitude is the effect of a discrete change from 0 to 1
in different civilizations dummy as a percentage of the probabilityof conflict at the means of the variables. All marginal effects are multiplied by
100 for readability. In addition to the variables displayed all columns include Contiguity, Log Geodesic Distance, Log Absolute Difference in Latitudes,
Log Absolute Difference in Longitudes, Number of Landlocked Countries in the Pair, Log Product of Land Areas in sq km, Ever in Colonial
Relationship, Common Colonizer, Countries were or are the Same Country, Different Legal Origins and Sum of Democracy Indexes.
20
promotes peace, i.e. the longer the countries have peaceful relations
with one another the less likely it is for them to be caught up in a …ght.
In columns (2) and (3) of Table 5 we account for the economic factors.
As such, we include the absolute di¤erence in log per capita income
and bilateral and multilateral trade relations to our regression. Observe
that the bigger the income di¤erences between countries the smaller
are the odds for a con‡ict. This might be due to the weaker country
acknowledging the fact of not being able to cope with a richer country
and looking for peaceful ways to settle the issues. To avoid reverse
causality between con‡ict and trade relations we lag the trade variables
by 4 periods. As in Martin et al. (2008) we establish the negative e¤ect
of bilateral interdependence and the positive e¤ect of trading with third
parties on con‡ict probability. In both cases distance operates in the
opposite direction and lessens the e¤ect of trade links.
4.2.2
Cold War and post-Cold War Comparisons
To carry Cold War and post-Cold War comparisons we start with looking
at Table 5 again. In column (4) of Table 5 we augment the speci…cation in column (3) by adding a post-Cold War dummy and its interaction
with di¤erent civilizations variable. This way we will have an idea about
the di¤erential e¤ects of civilizational dissimilarity in two di¤erent time
periods. Notice that the sign of post-Cold War dummy is negative which
tells us that in general the probability of con‡ict is smaller in the postCold War period. When we look at the interaction of post-Cold War
dummy and di¤erent civilizations variable we observe a positive and
signi…cant coe¢ cient which means that civilizational di¤erences matter
more in the post-Cold War world. Therefore, our results suggest that
while being part of di¤erent civilizations reduces the probability of con‡ict by 14.9 percentage points in the Cold War era, it increases the
likelihood of con‡ict by 86 percentage points in the post-Cold War era.
To further our analysis we split the sample into Cold War and postCold War periods. Results reporting the speci…cations with military
and economic factors in two di¤erent epochs are in Table 6.27 By looking at Table 6, although we observe that civilizational di¤erences are less
important in the post-Cold War period when we fail to account for eco27
A similar table to Table 6 for making Cold War and post-Cold War comparisons
controlling for geographic and political factors is reproduced in Table 6A in the
Appendix. Important to notice in Table 6A is that while the impact of di¤erent legal
origins weakens once we pass from Cold War to post-Cold War the e¤ect democracy
has increases. A possible explanation is that the e¤ect of di¤erent legal origins is
washed out in the post-Cold War period once more countries are democratic, hence
democracy has a larger impact and underlying di¤erent legal origins do not play such
a big role any more.
21
TABLE 6. Panel Regressions, probit. (Dependent variable: dichotomous indicator of
conflict between 1946-1991 and 1992-2001; Cold War and post-Cold War Comparisons).
(1)
(2)
(3)
(4)
military
military
economic
economic
factors,
factors,
factors,
factors,
Cold War
post-Cold War
Cold War post-Cold War
Different Civilizations
.0065**
.000378
-.00986
.0175***
(0.04)
(0.9)
(0.12)
(0.001)
.0578***
.0392***
.039***
.038***
(0.00)
(0.00)
(0.00)
(0.00)
.0041***
.0019
.0083***
-.00108
(0.00)
(0.17)
(0.00)
(0.63)
-.0133***
.0191**
-.002
-.004
(0.001)
(0.02)
(0.77)
(0.59)
.00024***
.000152*
.00038**
.000083
(0.003)
(0.08)
(0.013)
(0.47)
-.0033***
-.00199***
-.0038***
-.0017***
(0.00)
(0.00)
(0.00)
(0.00)
-.0101***
.00488**
(0.00)
(0.01)
Log Bilateral
-.0518***
-.02206**
Openness, t-4
(0.00)
(0.022)
Log Multilateral
.0136
.0838*
Openness, t-4
(0.65)
(0.07)
Log Distance×Log
.0071***
.00211*
Bilateral Openness
(0.00)
(0.071)
Log Distance×Log
-.0027
-.0127**
Multilateral Openness
(0.46)
(0.043)
Number of Major
Powers in the Pair
Log Absolute Difference
in Military Capabilities
Alliance Dummy
Number of Other
Wars in Year t
Number of
Peaceful Years
Absolute Difference
in Log per capita Income
# of Obs.
282061
118609
101463
48183
Standardized Magnitude(%)
21.972
1.536
-25.771
71.335
Robust p-values in parentheses; *** significant at 1%; ** significant at 5%; * significant at 10%. Probit marginal
effects are reported in all columns. For dummy variables,marginal effects are for discrete changes from 0 to 1. All
marginal effects are multiplied by 100 for readability. In addition to the variables displayed all columns include
Contiguity, Log Geodesic Distance, Log Absolute Difference in Latitudes, Log Absolute Difference in Longitudes,
Number of Landlocked Countries in the Pair, Log Product of Land Areas in sq km, Ever in Colonial Relationship,
Common Colonizer, Countries were or are the Same Country, Different Legal Origins and Sum of Democracy Indexes.
22
nomic factors, this conclusion gives way to another one when we control
for all of our variables on geographic, political, military and economic
factors. In columns (3)and (4) of Table 6 we show that when a pair
of countries in the post-Cold War era belong to di¤erent civilizations
they have 71.3 percentage points higher probability of being involved in
con‡ict than the countries that belong to the same civilization. During
the Cold War, on the other hand, more similar countries are prone to
militarized dispute and if two countries in a dyad are members of di¤erent civilizations their chances of con‡ict is reduced by 25.7 percentage
points. These results in columns (3) and (4) of Table 6 leads us to a conclusion that is supportive of Huntington’s thesis. Namely, civilizational
di¤erences are more nuanced in the post-Cold War era than in the Cold
War era.
5
Conclusion
As Huntington (1993b) said, faith and family, blood and belief, are what
people identify with and what they will …ght and die for. And that is why
the clash of civilizations is replacing the Cold War as the central phenomenon of global politics, and why a civilizational paradigm provides,
better than any alternative, a useful starting point for understanding
and coping with the changes going on in the world.
Though the above claim by Huntington might neglect several aspects
that feed into con‡ictual fault lines, there is an element of truth in it
as this study shows. By and large, our …ndings are supportive of Huntington’s hypothesis. We …nd that civilizational di¤erences do matter in
con‡ictual relations as they increase the likelihood of con‡ict by up to
62.8 percentage points over the period 1816-2011. More importantly,
civilizational di¤erences matter even more in the post-Cold War world
and country pairs that belong to di¤erent civilizations are associated
with 71.3 percentage points higher con‡ict probability than the ones
that belong to the same civilization.
23
A
Appendix
2001
1990
1980
1970
1960
1950
1940
1930
1920
1910
1900
1890
1880
1870
1860
1850
1840
1825
1816
0
Number of Conflict between 1816-2001
20
40
60
80
100
Figure 1A. Number of Con‡ict, 1816-2001
Source: Author’s Own Construction
2813 (64%)
1582 (36%)
0
Number of Conflict between 1816-2001
1,000
2,000
3,000
Figure 2A. Number of Con‡ict, 1816-2001
(Same Civilization and Di¤erent Civilizations Breakdown)
Same Civilization
Different Civilizations
Source: Author’s Own Construction
24
1301 (65.8%)
675 (34.1%)
0
Number of Conflict between 1946-1991
500
1,000
1,500
Figure 3A. Number of Con‡ict, 1946-1991
(Same Civilization and Di¤erent Civilizations Breakdown)
Same Civilization
Different Civilizations
Source: Author’s Own Construction
Figure 4A. Number of Con‡ict, 1991-2001
(Same Civilization and Di¤erent Civilizations Breakdown)
273 (39.6%)
0
Number of Conflict between 1991-2001
100
200
300
400
415 (60.3%)
Same Civilization
Different Civilizations
Source: Author’s Own Construction
25
TABLE 1A. Civilization Membership
Civilization Country
Western
Andorra, Australia,
Austria, Barbados,
Belgium, Canada,
Croatia, Czech Rep.,
Denmark, Dominica,
Estonia, Finland,
France, French Guiana,
Germany, Greenland,
Grenada, Hungary,
Iceland, Ireland,
Israel, Italy, Jamaica,
Latvia, Liechtenstein,
Lithuania, Luxembourg,
Malta, Monaco,
Netherlands,
New Zealand, Norway,
Papua New Guinea,
Philippines, Poland,
Portugal, San Marino,
Slovakia, Slovenia,
Solomon Islands, Spain,
Sweden, Switzerland,
Trinidad and Tobago,
United Kingdom,
United States, Vanuatu.
Sinic
China, Hong Kong,
North Korea,
South Korea, Taiwan,
Vietnam.
Islamic
Afghanistan, Albania,
Algeria, Azerbaijan,
Bahrain, Bangladesh,
Bosnia and Herzegovina,
Brunei, Burkina Faso,
Chad, Djibouti,
26
TABLE 1A. Continued.
Civilization
Country
Islamic
Egypt, Eritrea,
Gambia, Guinea,
Guinea-Bissau,
Indonesia, Iran,
Iraq, Jordan,
Kyrgyzstan, Kuwait,
Lebanon, Libya,
Malaysia, Mali,
Mauritania, Morocco,
Niger, Oman,
Pakistan, Qatar,
Saudi Arabia, Senegal,
Somalia, Sudan, Syria,
Tajikistan, Tunisia,
Turkey, Turkmenistan,
United Arab Emirates,
Uzbekistan, Yemen.
Hindu
Guyana, India,
Nepal.
Orthodox
Armenia, Belarus,
Bulgaria, Cyprus,
Georgia, Greece,
Kazakhstan, Macedonia,
Moldova, Romania,
Russia, Serbia, Ukraine.
Latin American Antigua and Barbuda,
Argentina, Bahamas,
Belize, Bolivia,
Brazil, Chile,
Colombia, Costa Rica,
Cuba, Dominican Rep.,
Ecuador, El Salvador,
Guatemala, Honduras,
Mexico, Nicaragua,
Panama, Paraguay,
27
TABLE 1A. Continued.
Civilization
Country
Latin American Peru, Puerto Rico,
Saint Lucia,
St.Vincent & Grenadines,
Uruguay, Venezuela.
African
Angola, Benin,
Botswana, Burundi,
Cameroon, Cape Verde,
Central African Republic,
Comoros, Congo,
Congo, Dem. Rep. (Zaire),
Equatorial Guinea,
Gabon, Ghana,
Ivory Coast, Kenya,
Lesotho, Liberia,
Madagascar, Malawi,
Mauritius, Mozambique,
Namibia, Nigeria, Rwanda,
Sao Tome and Principe,
Sierra Leone,
South Africa, Suriname,
Swaziland, Tanzania,
Togo, Uganda,
Zambia, Zimbabwe.
Buddhist
Bhutan, Cambodia,
Lao People’s Dem. Rep.,
Mongolia, Myanmar,
Singapore, Sri Lanka,
Thailand.
"Lone" States
Ethiopia, Haiti, Japan.
Source: Author’s own construction
based on Huntington (1998).
28
TABLE 2A. Number of Conflicts within and across Civilizations between 1946-1991.
Civilizations
Western
Sinic
Islamic
Hindu
Orthodox Latin American African
Buddhist Lone States
Western
35 (1.7%)
Sinic
200 (10%)
Islamic
387 (19.5)
14 (.7%)
241 (12.1%)
4 (.2%)
33 (1.6%)
52 (2.6%)
Orthodox
130 (6.5%) 41 (2.0%)
56 (2.8%)
0
14 (.7%)
Latin American
67 (3.3%)
7 (.35%)
8 (.4%)
10 (.5%)
5 (.25%)
90 (4.5%)
African
29 (1.4%)
0
32 (1.6%)
3 (.15%)
2 (.1%)
21 (1%)
Buddhist
11 (.55%)
77 (3.8%)
2 (.1%)
5 (.25)
5 (.25%)
0
0
68 (3.4)
Lone States
5 (.25%)
28 (1.4%)
33 (1.6%)
0
27 (1.3%)
7 (.35%)
0
0
Hindu
97 (4.9%)
3 (.15%)
127 (6.4%)
0
Source: Author’s own construction
TABLE 3A. Number of Conflicts within and across Civilizations between 1991-2001.
Civilizations
Western Sinic
Western
10 (1.4%)
Sinic
30 (4.3%)
30 (4.3%)
Islamic
65 (9.4%)
1 (.14%)
Hindu
Orthodox
Latin American
African
Buddhist
Lone States
Islamic
Hindu
Orthodox Latin American African
Buddhist Lone States
98 (14.2%)
1 (.14%)
0
13 (1.8%)
0
138 (20%)
2 (.29%)
87 (12.6%)
0
31 (4.5%)
7 (1%)
0
0
1 (.14%)
2 (.29%)
28 (4%)
2 (.29%)
0
21 (3%)
2 (.29%)
0
0
66 (9.5%)
0
3 (.43%)
1 (.14%)
1 (.14%)
0
0
0
10 (1.4%)
10 (1.4%)
13 (1.8%)
5 (.72%)
0
6 (.87%)
3 (.43%)
1 (.14%)
0
Source: Author’s own construction
29
0
TABLE 4A. Summary Statistics
Variable
# Obs. Mean Std. Dev. Min
Max
Con‡ict (%)
Di¤erent Civilizations
Contiguity Dummy
Log Distance
Log Absolute Latitude Di¤erence
Log Absolute Longitude Di¤erence
# Landlocked Countries in a Dyad
Log Product of Land Area
Colonial Relationship Dummy
Common Colonizer Dummy
Part of Same Polity Dummy
Di¤erent Legal Origins Dummy
Sum of Democracy Indexes
# Major Powers in a Dyad
Log Absolute CINC Di¤erence
Alliance Dummy
# Other Wars in Year t
# Peaceful Years
Absolute Log Income Di¤erence
Log Bilateral Openness
Log Multilateral Openness
100
1
1
9.89
4.64
5.61
2
32.76
1
1
1
1
20
2
-.957
1
107
186
12.26
2.13
8.95
590337
590337
590337
589696
586598
586818
586845
586845
599328
501041
593424
613318
490839
608431
589315
613297
607076
590337
278897
225272
225448
30
.744
.807
.046
8.714
2.903
3.735
.347
24.30
.0159
.094
.0107
.632
-.3869
.108
-5.944
.057
22.57
27.89
2.812
-6.887
-1.387
8.596
.394
.210
.8009
1.135
1.21
.535
3.077
.125
.291
.103
.482
10.6
.323
2.02
.232
19.05
27.32
2.039
2.27
.848
0
0
0
4.29
-4.09
-4.60
0
9.86
0
0
0
0
-20
0
-18.42
0
0
0
.0000095
-16.78
-10.41
TABLE 5A. Cross-Sectional Regressions, probit.
(Dependent variable: dichotomous indicator for whether a country
pair was ever involved in a war, Cold War and post-Cold War Comparison).
(1)
(2)
(3)
(4)
Cold War post-Cold War Cold War post-Cold War
period
period
period
period
.096***
.0046
.076***
.0042
(0.002)
(0.78)
(0.004)
(0.79)
.3509***
.0094
.313***
.0066
(0.002)
(0.74)
(0.002)
(0.8)
-.0213
-.0142
-.0202
-.012
(0.42)
(0.26)
(0.37)
(0.3)
-.055***
-.0103*
-.048***
-.009
in Latitudes
(0.00)
(0.09)
(0.00)
(0.11)
Log Absolute Difference
-.0187
-.0084
-.015
-.008
in Longitudes
(0.23)
(0.17)
(0.23)
(0.17)
-.216***
.0356***
-.181***
.034***
Countries in the Pair
(0.00)
(0.001)
(0.00)
(0.001)
Log Product of Land
.042***
.0125***
.035***
.0122***
Areas in sq km
(0.00)
(0.00)
(0.00)
(0.00)
Different Civilizations
Contiguity
Log Geodesic Distance
Log Absolute Difference
Number of Landlocked
Ever in Colonial Relationship
-.045
(0.48)
Common Colonizer
-.0627
-.0092
(0.15)
(0.64)
Countries were or are
.398*
.0408
the Same Country
(0.07)
(0.45)
Different Legal Origins
.077***
-.011
(0.004)
(0.46)
# Obs.
Standardized magnitude(%)
15309
15309
15309
15136
111.219
19.554
107.496
18.821
p-values in parentheses; *** significant at 1%; ** significant at 5%; * significant at 10%.
Probit marginal effects are reported in all columns. For dummy variables, marginal effects are for
discrete changes.from 0 to 1. The standardized magnitude is the effect of a discrete change
from 0 to 1 in different civilizations dummy as a percentage of the probability of war at the
means of the variables. All marginal effects are multiplied by 100 for the sake of readability.
31
TABLE 6A. Panel Regressions, probit. (Dependent variable: dichotomous indicator of
conflict between1946-1991 and 1992-2001; Cold War and post-Cold War Comparisons).
(1)
(2)
(3)
(4)
geographic
geographic
political
political
factors,
factors,
factors,
Cold War
Different Civilizations
Contiguity
Log Geodesic Distance
Log Absolute Difference
in Latitudes
post-Cold War Cold War
factors,
post-Cold War
.0842***
.0522***
.0692***
.0163
(0.00)
(0.00)
(0.00)
(0.14)
1.286***
.333***
1.177***
.802***
(0.00)
(0.00)
(0.00)
(0.00)
-.05***
-.0626***
-.0436***
-.0046
(0.00)
(0.00)
(0.00)
(0.6)
-.061***
-.0528***
-.0724***
-.0508***
(0.00)
(0.00)
(0.00)
(0.00)
-.0487***
-.038***
-.0638***
-.0408***
(0.00)
(0.00)
(0.00)
(0.00)
-.144***
-.0615***
-.1486***
-.046***
Countries in the Pair
(0.00)
(0.00)
(0.00)
(0.00)
Log Product of Land
.0548***
.0318***
.0554***
.0288***
(0.00)
(0.00)
(0.00)
(0.00)
.0545*
.2439***
(0.08)
(0.00)
-.0372**
-.0423***
(0.02)
(0.001)
.013
.1968***
the Same Country
(0.58)
(0.00)
Different Legal Origins
.106***
.0182*
Log Absolute Difference
in Longitudes
Number of Landlocked
Areas in sq km
Ever in Colonial Relationship
Common Colonizer
Countries were or are
Sum of Democracy Indexes
(0.00)
(0.06)
-.00058*
-.0034***
(0.10)
(0.00)
# of Obs.
333033
152312
282067
118614
Standardized Magnitude(%)
72.181
70.046
55.513
23.542
Robust p-values in parentheses; *** significant at 1%; ** significant at 5%; * significant at 10%.
Probit marginal effects are reported in all columns. For dummy variables,marginal effects are for
discrete changes from 0 to 1. The standardized magnitude is the effect of a discrete
change from 0 to 1 in different civilizations dummy as a percentage of the probability of conflict
at the means of the variables. All marginal effects are multiplied by 100 for readability.
32
References
[1] Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S. and
Wacziarg, R. 2003. "Fractionalization." Journal of Economic
Growth, 8, 155-194.
[2] Barbieri, K., Keshk, O. and Pollins, B. 2008. "Correlates of War
Project Trade Data Set Codebook, Version 2.0."
[3] Barbieri, K., Keshk, O. and Pollins, B. 2009. “TRADING DATA:
Evaluating our Assumptions and Coding Rules.”Con‡ict Management and Peace Science, Forthcoming.
[4] Bolks, S. and Stoll, R. 2003. "Examining Con‡ict Escalation Within
the Civilizations Context." Con‡ict Management and Peace Science
20(2): 85–110.
[5] Bremer, S. A. 1992. "Dangerous Dyads: Conditions A¤ecting the
Likelihood of Interstate War, 1816–1965." Journal of Con‡ict Resolution 36(2): 309–341.
[6] Chiozza, G. 2002. "Is There a Clash of Civilizations? Evidence from
Patterns of International Con‡ict Involvement, 1946–97." Journal
of Peace Research 39: 711–34.
[7] Correlates of War Project. 2003. "Alliances Data, 1816-2000 Version 3.03."
[8] Correlates of War Project. 2007. "Direct Contiguity Data, 18162006. Version 3.1."
[9] Correlates of War Project. 2007. "Militarized Interstate Disputes
Data, 1816-2001. Version 3.1."
[10] Correlates of War Project. 2007. "National Material Capabilities
Data, 1816-2007. Version 4.0."
[11] Correlates of War Project. 2008. “State System Membership List,
1816-2004. Version 2008.1.”
[12] Correlates of War Project. 2008. "Trade Data Set, 1870-2006. Version 2.0."
[13] Correlates of War Project. 2011. "COW Wars Data, 1816-2007. Version 4.0."
[14] Easterly, W. and Levine, R. 1997. "Africa’s Growth Tragedy: Policies and Ethnic Divisions." The Quarterly Journal of Economics,
Vol. 112, No. 4 (Nov.), 1203-1250.
[15] Fearon, J. D. 2003. "Ethnic and Cultural Diversity by Country."
Journal of Economic Growth, 8, 195-222.
[16] Fearon, J. and Laitin, D. 2003. "Ethnicity, Insurgency, and Civil
War." American Political Science Review, 97(1): 75-90.
[17] Fukuyama, F. 1992. "The End of History and the Last Man." Free
Press.
[18] Garcia-Montalvo, J. and Reynal-Querol, M. 2002. "Why Ethnic
33
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
Fractionalization? Polarization, Ethnic Con‡ict and Growth."
Mimeo, Universitat Pompeu Fabra, September.
Gar…nkel, M. R. and Skaperdas, S. 2006. "Economics of Con‡ict:
An Overview." in K. Hartley and T. Sandler (eds) Handbook of
Defense Economics, Volume 2, North Holland, Amsterdam.
Gartzke, E. and Gleditsch, K. S. 2006. “Identity and Con‡ict: Ties
that Bind and Di¤erences that Divide.”European Journal of International Relations, 12(1): 53-87.
Ghosn, F. and Bennett, S. 2003. "Codebook for the Dyadic Militarized Interstate Incident Data, Version 3.10."
Gibler, D. M. 2009. "International Military Alliances from 1648 to
2008." Washington: Congressional Quarterly Press.
Gibler, D. M. and Sarkees, M. 2004. “Measuring Alliances: The
Correlates of War Formal Interstate Alliance Data set, 1816-2000.”
Journal of Peace Research 41(2): 211-222.
Ghosn, F., Palmer, G. and Bremer, S. 2004. "The MID3 Data Set,
1993–2001: Procedures, Coding Rules, and Description." Con‡ict
Management and Peace Science 21:133-154.
Gleditsch, N. P. and Singer, J. D. 1975. "Distance and International
War, 1816-1965." in Proceedings of the International Peace Research Association (IPRA) Fifth General Conference (Oslo: IPRA):
481-506.
Gochman, C. S. 1991. "Interstate Metrics: Conceptualizing, Operationalizing, and Measuring the Geographic Proximity of States
since the Congress of Vienna." International Interactions 17 (1):
93-112.
Head, K. and Mayer, T. 2002. "Illusory Border E¤ects: Distance
Mismeasurement In‡ates Estimates of Home Bias in Trade." CEPII,
Working Paper No 2002-01.
Henderson, E., A. 1997. "Culture or Contiguity? Ethnic Con‡ict,
the Similarity of States, and the Onset of War, 1820–1989." Journal
of Con‡ict Resolution 41(5): 649–68.
Henderson, E., A. 1998. "The Democratic Peace through the Lens
of Culture, 1820–1989." International Studies Quarterly 42(3): 461–
84.
Henderson, E., A. and Tucker, R., M. 2001. "Clear and Present
Strangers: The Clash of Civilizations and International Con‡ict."
International Studies Quarterly 45: 317–38.
Huntington, S., P. 1993a. "The Clash of Civilizations?." Foreign
A¤airs 72(3): 22-49.
Huntington, S., P. 1993b. "If Not Civilizations, What? Paradigms
of the Post-Cold War World." Foreign A¤airs 72(5): 186–94.
34
[33] Huntington, S., P. 1998. "The Clash of Civilizations and the Remaking of the World Order." Simon & Schuster Ltd. West Garden
Place Kendal Street London W2 2AQ.
[34] Huntington, S., P. 2000. "Try Again: A Reply to Russett, Oneal
and Cox." Journal of Peace Research 37(5): 609–10.
[35] Jakobsen, J. and Jakobsen, T., G. 2010. "Birds of a Feather Flock
Apart? Testing the Critique of the Clash of Civilizations Thesis."
Journal of Peace, Con‡ict and Development, Issue.15.
[36] La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R. 1999.
"The Quality of Government." Journal of Law, Economics and Organization 15(1), 222-279.
[37] Levy, G. and Razin, R. 2004. “It Takes Two: An Explanation for the
Democratic Peace.”Journal of the European Economic Association,
2 (1), 1–29.
[38] Martin, P., Mayer, T. and Thoening, M. 2008. "Make Trade Not
War?" Review of Economic Studies, 75: 865-900.
[39] Mauro, P. 1995. "Corruption and Growth." The Quarterly Journal
of Economics, Vol. 110, No. 3 (Aug.), 681-712.
[40] Mungiu-Pippidi, A. and Mindruta, D. 2002. "Was Huntington
Right? Testing Cultural Legacies and the Civilization Border." International Politics, Vol. 39, No. 2, pp. 193-213.
[41] Norris, P. and Inglehart, R. 2002. "Islamic Culture and Democracy:
Testing the Clash of Civilizations Thesis." Comparative Sociology,
1(3): 235-263.
[42] Russett, B., M., Oneal, J., R. and Cox, M. 2000. "Clash of Civilizations, or Realism and Liberalism D‘ej´a Vu?" Journal of Peace
Research, 37(5): 583–608.
[43] Sarkees, Meredith Reid and Wayman, Frank. 2010. "Resort to War:
1816 - 2007." CQ Press.
[44] Small, M. and Singer, J., D. 1969. "Formal Alliances, 1815-1965: An
Extension of the Basic Data." Journal of Peace Research 6:257-282.
[45] Singer, J. D. 1987. "Reconstructing the Correlates of War Dataset
on Material Capabilities of States, 1816-1985." International Interactions, 14: 115-32.
[46] Singer, J., D. and Small, M. 1966. "Formal Alliances, 1815-1939."
Journal of Peace Research 3:1-31.
[47] Singer, J., D., Bremer, S. and Stuckey, J. 1972. "Capability Distribution, Uncertainty, and Major Power War, 1820-1965." in Bruce
Russett (ed) Peace, War, and Numbers, Beverly Hills: Sage, 19-48.
[48] Spolaore, E. and Wacziarg, R. 2010. "War and Relatedness."
Mimeo.
[49] Stinnett, Douglass, M., Jaroslav Tir, Philip Schafer, Paul F. Diehl,
35
and Charles Gochman. 2002. "The Correlates of War Project Direct Contiguity Data, Version 3." Con‡ict Management and Peace
Science 19 (2):58-66.
[50] Treisman, D. 2000. "The Causes of Corruption: A Cross-national
Study." Journal of Public Economics, 76 (3), 399–457.
36